🔊Sound Design Unit 16 – Audio Post-Production: Stems and Delivery

What Are Stems?

• Stems are submixes of audio elements (dialogue, music, sound effects) that are rendered separately for flexibility in the final mix
• Typically consist of a group of related sounds or instruments mixed together (all drum elements, all lead vocals, all background ambiences)
• Allow for more control and adjustments to specific elements without affecting the entire mix
• Provide a way to deliver separated audio elements to clients or other professionals for further mixing or localization
• Commonly used in film, television, and video game audio post-production workflows
  • Enable efficient collaboration between sound designers, composers, and re-recording mixers
  • Facilitate the creation of different language versions or platform-specific mixes (theatrical, streaming, broadcast)

Why Use Stems in Audio Post-Production?

• Enable non-destructive changes to the mix without requiring a complete remix from individual tracks
• Allow for efficient delivery of audio elements to other professionals or clients for further mixing, localization, or implementation
• Provide flexibility to create different versions of the mix (theatrical, broadcast, streaming) without starting from scratch
• Facilitate the creation of different language versions of the audio mix by allowing dialogue replacement without affecting music and sound effects
• Streamline the process of implementing audio into interactive media (video games) by delivering separated audio elements
• Improve collaboration between sound designers, composers, and re-recording mixers by allowing each to focus on their specific area of expertise
• Reduce the risk of errors or unintended changes to the mix by limiting the scope of adjustments to specific stems

Types of Stems in Sound Design

• Dialogue Stems: Contain all spoken words, narration, and vocalized sounds from characters
  • Can be further separated into specific characters, languages, or types of dialogue (on-screen, off-screen, voiceover)
• Music Stems: Consist of the musical elements of the soundtrack
  • Often split into subgroups (lead instruments, background elements, percussion, vocals)
  • Can be delivered as full mixes or separate stems for each instrument or group
• Sound Effects Stems: Include all non-musical and non-dialogue sounds
  • Often organized by type (hard effects, foley, ambiences, vehicles, weapons)
  • Can be split into more specific categories (creature vocalizations, magic spells, user interface sounds)
• Ambience and Background Stems: Contain environmental sounds and background noise
  • Can include room tones, outdoor ambiences, crowd sounds, and other non-specific atmospheric elements
• Foley Stems: Consist of synchronized, post-recorded sound effects
  • Often used for footsteps, cloth movement, prop handling, and other detailed sounds
• Specialized Stems: Can be created for specific purposes or elements
  • Examples include stems for interactive elements in video games (menu sounds, player feedback) or separate stems for immersive audio formats (Dolby Atmos, DTS:X)

Creating and Organizing Stems

• Start by identifying the key elements of the audio mix (dialogue, music, effects) and determining the necessary level of separation
• Create a consistent naming convention for stems to ensure clarity and organization
  • Use descriptive names that clearly identify the content of each stem (Dialogue_Main_Characters, Music_Background, SFX_Ambiences)
  • Include version numbers or dates to keep track of revisions
• Use audio editing software (Pro Tools, Nuendo, Reaper) to create and edit individual stems
  • Route related tracks to specific output buses or tracks for rendering
  • Apply processing (EQ, compression, effects) to each stem as needed
• Render each stem as a separate audio file, typically in a lossless format (WAV, AIFF) at the required sample rate and bit depth
• Organize rendered stems in a clear folder structure
  • Use subfolders for each category (Dialogue, Music, SFX) and further subfolders for more specific elements if needed
• Create a master document or spreadsheet that lists all stems, their contents, and any relevant notes or instructions for mixing or implementation

Mixing with Stems

• Import stems into your mixing session, maintaining the original organization and naming convention
• Create submixes for each category of stems (Dialogue, Music, SFX) to allow for broad adjustments
• Use routing and bussing to apply processing (EQ, compression, reverb) to specific stems or groups of stems
  • Avoid applying excessive processing to individual stems, as this can limit flexibility in the final mix
• Balance the levels of each stem within its category submix, ensuring clarity and cohesion
• Create a master output bus for the final mix and adjust the levels of each category submix to achieve the desired balance
• Use automation to create dynamic changes in the mix over time
  • Ride the levels of specific stems or submixes to emphasize important elements or create space for others
• Regularly reference the mix on different playback systems (speakers, headphones) to ensure translation and compatibility
• Create alternate versions of the mix (stereo, surround, immersive) by routing stems to the appropriate output channels
• Render the final mix as a master file, typically in a lossless format (WAV, AIFF) at the required sample rate and bit depth

Stem Delivery Formats and Standards

• Stems are typically delivered as individual audio files in a lossless format (WAV, AIFF) to maintain quality
• Sample rate and bit depth should match the project requirements, often 48kHz/24-bit for film and television
• File names should follow a consistent convention and include relevant information (project name, stem name, version number)
• Stems should be synchronized to a common timecode or start time to ensure proper alignment in the final mix
• Deliver stems in a organized folder structure with clear labeling and documentation
  • Include a master document or spreadsheet that lists all stems, their contents, and any relevant notes or instructions
• Conform to any specific delivery requirements or standards provided by the client or platform
  • Some platforms may have specific file format, naming, or metadata requirements
• Consider the intended use of the stems and provide additional information or alternate formats if necessary
  • For example, video game audio may require separate files for each individual sound or loop points for seamless playback
• Use archival-quality media (hard drives, SSDs) for physical delivery or secure file transfer protocols (FTP, Aspera) for digital delivery

Quality Control and Troubleshooting

• Thoroughly review each stem before delivery to ensure quality and completeness
  • Listen for any unwanted noise, distortion, or artifacts and address them in the source files
  • Check for proper synchronization and alignment of stems to the project timeline
• Verify that all necessary stems are included in the delivery package and that file names and formats match the project requirements
• Test the stems in the intended playback environment (mixing software, game engine) to ensure compatibility and proper functionality
• Address any issues or errors found during quality control checks
  • Repair or replace problematic files, update documentation, or communicate with the client or team for guidance
• Maintain detailed notes and documentation of any changes or fixes made during the quality control process
• Perform a final check of the delivered package to ensure all files and documentation are included and properly organized
• Provide contact information and support for any issues or questions that may arise after delivery
• Establish a versioning system and keep backups of all files to allow for easy retrieval and updates if needed

Future Trends in Stem-Based Audio Post-Production

• Increased adoption of immersive audio formats (Dolby Atmos, DTS:X) may require more complex stem creation and delivery
  • Additional stems for height channels, object-based audio, or specific speaker configurations
• Advances in machine learning and AI may streamline the process of creating and organizing stems
  • Automated dialogue isolation, music separation, and sound effect categorization
• Cloud-based collaboration platforms may enable more efficient remote work and real-time updates to stem files
• Integration of virtual and augmented reality technologies may require new approaches to stem creation and implementation
  • Spatial audio, interactive sound design, and real-time mixing based on user position and interactions
• Expansion of interactive and non-linear media (video games, VR experiences) may necessitate more flexible and modular stem architectures
  • Stems designed for dynamic mixing, procedural audio generation, and real-time effects processing
• Continued growth of localization and accessibility requirements may drive the need for more detailed and separated stem deliveries
  • Separate stems for each language, descriptive audio tracks, or customizable mix options for end-users